1.Install python 3.5, keras 2.1.3, tensorflow 1.13.1, and Matlab 2019a (2h is needed to intall all softwares) 2.Download the all the files from https://github.com/lungproject/lungegfr, and the trained model (for EGFR prediction: lungEGFR.hdf5; for PD-L1 prediction: lungIO.hdf5) was in the folder named model. 3.Run GenerateNPY.m with Matlab to obtain all the sample .npy file from the .png file 4.Run Code- test_samplepatient.py with python to generate the prediction results of sample data in the folder named data. 5.Downlad all the data from https://pan.baidu.com/s/1YpxlrZf5WSqrTLg2h4Fbog (code: om9i) 6.Run test_allpatient.py with python to generate the prediction results of all the input slices. (35 seconds are needed to run all the data) 7.Run Obtainresults.m with Matlab to obtain the predicted DLS of each patient. (25 seconds are needed to obtain the final results)